This proposal describes the development of new computational algorithms and potential functions for determining protein structure from the sequence and limited structural information. Pr,eliminary results demonstrate that low resolution (approximately 6A ) structures can be obtained for a complicated protein such as myoglobin using a reduced model of the protein if secondary structure is specified; results for several other helical proteins and one mixed alpha/beta protein have also been obtained. Detailed, all-atom structures have been generated from these reduced model structures via input into a molecular mechanics program, addition of side chalns, and minimization and/or simulated annealing. The generalize Born continuum solvent model of Still and coworkers or numerical solution of the Poisson-Boltzmann equation is used to treat solvent effects. Quantum chemical reaction field methods will be used to develop new high resolution potential functions. The initial goal of the proposal is to further develop this methodology so that, given secondary structure, 3-4A structures can reliably be obtained for an arbitrary protein from the reduced model, and 1-2 A structures can then be generated at the molecular mechanics level of representation. This technology can then be used to qualitatively extend the range of proteins amenable to,NMR structure determination and reduce the time to solution, as the number of long range distance constraints per residue that are required will be diminished considerably. On a longer timescale, there are prospects for developing new experimental methods, involving molecular biology and optical spectroscopy, for studying proteins inaccessible to NMR. This is a more speculative endeavor and will require novel, as yet undetermined advances in experimental methodology as well as improved computational algorithms. Achievement of the practical goals of more effective structure determination will have considerable impact on basic structural biology and on rational drug design efforts. On a more fundamental level, studies of a large number of protein conformations and potential functions will yield substantial insight into the physical chemistry of protein folding and the construction of protein models which accurately describe this chemistry.